Substrate envelope-designed potent HIV-1 protease inhibitors to avoid drug resistance

Article abstract of DOI:10.1016/j.chembiol.2013.07.014

Summary The rapid evolution of HIV under selective drug pressure has led to multidrug resistant (MDR) strains that evade standard therapies. We designed highly potent HIV-1 protease inhibitors (PIs) using the substrate envelope model, which confines inhib

Full text of DOI:10.1016/j.chembiol.2013.07.014

Chemistry & Biology
Article
Substrate Envelope-Designed Potent HIV-1
Protease Inhibitors to Avoid Drug Resistance
Madhavi N.L. Nalam,1 Akbar Ali,1 G.S. Kiran Kumar Reddy,1 Hong Cao,1 Saima G. Anjum,1,6 Michael D. Altman,2,3,5
1,4,
1,
Nese Kurt Yilmaz,1 Bruce Tidor,2,3, Tariq M. Rana,
*
*
*
and Celia A. Schiffer
1Department of Biochemistry and Molecular Pharmacology, University of Massachusetts Medical School, Worcester, MA 01605, USA
2Department of Biological Engineering
3Department of Electrical Engineering and Computer Science
Massachusetts Institute of Technology, Cambridge, MA 02139, USA
4Present address: Program for RNA Biology, Sanford-Burnham Medical Research Institute, 10901 North Torrey Pines Road, La Jolla,
CA 92037, USA
5Present address: Merck Research Laboratories, 33 Avenue Louis Pasteur, Boston, MA 02115, USA
6Present address: Department of Biological Sciences, University of Massachusetts Boston, 100 Morrissey Boulevard, Boston,
MA 02125, USA
*Correspondence: tidor@mit.edu (B.T.), trana@sanfordburnham.org (T.M.R.), celia.schiffer@umassmed.edu (C.A.S.)
http://dx.doi.org/10.1016/j.chembiol.2013.07.014
SUMMARY
tase promote the emergence of drug-resistant viral strains in
patients undergoing therapy. New patients are also infected
The rapid evolution of HIV under selective drug pres- with already resistant viruses, which is an added challenge in
the treatment of HIV infection. Thus, novel potent drugs targeting
sure has led to multidrug resistant (MDR) strains that
the drug-resistant viral ensemble are needed for effective treat-
ment of HIV/AIDS.
evade standard therapies. We designed highly
potent HIV-1 protease inhibitors (PIs) using the sub-
strate envelope model, which confines inhibitors
within the consensus volume of natural substrates,
providing inhibitors less susceptible to resistance
because a mutation affecting such inhibitors will
Various strategies have been used to develop new antiviral PI
therapies against drug-resistant HIV, including increasing the
plasma levels of existing PIs by using a boosting agent (Kempf
et al., 1997; Youle, 2007; Zeldin and Petruschke, 2004) and
developing new PIs using structure-based drug design (Ghosh
simultaneously affect viral substrate processing.
The designed PIs share a common chemical scaffold
but utilize various moieties that optimally fill the sub-
et al., 2008; Gulnik and Eissenstat, 2008; Nalam and Schiffer,
2008; Wensing et al., 2010). The design strategy to maximize
the number of hydrogen bonds with the protease backbone led
strate envelope, as confirmed by crystal structures. to the development of highly potent PIs active against drug-resis-
tant HIV (Ghosh et al., 2008, 2009, 2011). PIs with improved resis-
tance profiles were also developed using a solvent anchoring
approach (Cihlar et al., 2006) and utilizing a new lysine sulfon-
The designed PIs retain robust binding to MDR
protease variants and display exceptional antiviral
potencies against different clades of HIV as well as
a panel of 12 drug-resistant viral strains. The sub-
strate envelope model proves to be a powerful strat-
egy to develop potent and robust inhibitors that
avoid drug resistance.
amide-based molecular core (Stranix et al., 2003). A more
comprehensive strategy is to incorporate the substrate envelope
constraints in the structure-based drug design. This strategy is
based on the observation that substrates with high sequence
diversity adopt a conserved shape when bound to HIV-1 prote-
ase (Prabu-Jeyabalan et al., 2002). The consensus volume
shared by the bound substrates defines the substrate envelope.
INTRODUCTION
Drug-resistance mutations often occur at sites where the inhibi-
tor protrudes outside the substrate envelope (King et al., 2004a).
Drug resistance is a major problem in the treatment of patients Similar observations have recently been demonstrated to be
with HIV/AIDS. Resistance in HIV occurs when the target enzyme valid for the HCV NS3/4A protease as well (Romano et al.,
mutates, leading to less efficient drug binding, but maintains bio- 2010, 2012), supporting the generality of the substrate envelope
logic function. Currently, there are 25 US Food and Drug Admin- model. Inhibitors designed to fit within this envelope are likely to
istration (FDA)-approved drugs targeting different stages in the be less susceptible to drug resistance (http://hivdb.Stanford.
life cycle of HIV, including nine protease inhibitors (PIs; http:// edu) because a mutation that impairs inhibitor binding will simul-
www.fda.gov/ForConsumers/ByAudience/ForPatientAdvocates/ taneously affect substrate recognition and protease function.
HIVandAIDSActivities/ucm118915.htm. These drugs, especially
Previously, we designed and synthesized series of PIs with
when used in combination as the highly active antiretroviral and without substrate envelope constraints (Ali et al., 2006,
therapy (HAART), have improved the quality and life expectancy 2010; Altman et al., 2008; Chellappan et al., 2007; Nalam et al.,
of patients with HIV infection (Hogg et al., 1998; Palella et al., 2010; Nalam and Schiffer, 2008). These studies revealed that
1998). However, the very high replication rate of the virus and in contrast to inhibitors designed without any constraint, inhibi-
the lack of an error-proof mechanism in HIV reverse transcrip- tors designed to fit within the substrate envelope retain binding
1116 Chemistry & Biology 20, 1116–1124, September 19, 2013 ª2013 Elsevier Ltd All rights reserved

Chemistry & Biology
HIV-1 Protease Inhibitors to Avoid Drug Resistance
the protease due to enhanced size and flexibility, while still stay-
ing within the substrate envelope (Altman et al., 2008; Nalam
et al., 2010). The challenge is combining these moieties into a
single inhibitor that retains extremely high potency and stays
within the volume of the substrate envelope.
We used the substrate envelope strategy to optimize PIs
based on the DRV scaffold by incorporating ligands at the P10
and P20 positions that fit within the substrate envelope and there-
fore are predicted as promising candidates to be potent inhibi-
tors of multidrug-resistant (MDR) HIV-1 protease variants. In
this study, we report the structure-guided design and synthesis
of a series of such potent inhibitors and their subsequent struc-
tural, biochemical, antiviral, and pharmacokinetic evaluation.
Two chemical moieties at the P10 position and five chemical moi-
eties at the P20 position were incorporated to optimize inhibitor
flexibility and interactions with the protease. The resulting ten
inhibitors (Figure 1) were evaluated for their inhibitory activity
against wild-type (WT) and MDR HIV-1 protease variants, anti-
viral activity against a panel of WT and patient-derived drug
resistant HIV strains, and pharmacokinetic properties. The com-
pounds retain highly potent inhibitory activity against MDR pro-
tease variants, display exceptional potency against a panel of
12 MDR strains in antiviral assays, and overall perform better
than the most potent FDA-approved inhibitor, DRV. Crystal
structures of complexes with WT HIV-1 protease reveal that
the inhibitors indeed fit within the substrate envelope as
designed and support the use of substrate envelope constraints
in the design of potent inhibitors evading resistance.
RESULTS
Inhibitor Design and Synthesis
The design strategy for high-affinity inhibitors against drug-resis-
tant HIV-1 protease variants was based on three criteria: The
inhibitors should (1) fit within the substrate envelope; (2) be
based on the scaffold of a known high-affinity inhibitor; and (3)
have optimized P10 and P20 groups to retain contacts with the
protease with drug-resistance mutations.
Figure 1. Structures of the Designed Protease Inhibitors
(A) The chemical scaffold and P10 and P20 groups of the ten inhibitors.
(B) Superimposed X-ray crystal structures of all ten inhibitors in complex with
WT HIV-1 protease. See also Table S1 for crystallographic statistics.
(C) Fit of inhibitors within the substrate envelope. The substrate envelope is in
blue space filling representation, and the superimposed inhibitors are dis-
played as sticks. Protrusions beyond the substrate envelope are in red.
The primary constraint in the design of inhibitors was proper fit
within the substrate envelope of HIV-1 protease. This approach
aims to minimize susceptibility to resistance because a mutation
that affects inhibitor binding will simultaneously affect substrate
affinity (King et al., 2004a). In accord with the second criterion,
the design was based on the (R)-(hydroxyethylamino)sulfon-
amide dipeptide isostere, the shared core scaffold of APV and
affinity to drug-resistant HIV-1 protease variants, thereby having DRV. Finally, different chemical moieties were introduced at
flatter resistance profiles. A careful examination of darunavir the P10 and P20 positions. As the isobutyl P10 moiety in DRV loses
(DRV) binding to HIV-1 protease by free energy analysis sug- van der Waals contacts with the protease variants containing
gested that the interactions in the S10 and S20 sites can be drug-resistance mutations I50V, V82A, and I84V (King et al.,
improved by using more hydrophobic P10 and P20 groups to sus- 2004b), two P10 ligands, (S)-2-methylbutyl (isopentyl) and
tain van der Waals contacts to protease variants with active site 2-ethyl-n-butyl (isohexyl), with enhanced flexibility and steric
mutations (Cai and Schiffer, 2010). In previous structure-activity volume were probed to enable sustained contacts with the pro-
relationship (SAR) studies of PI libraries based on the (R)- tease. The combination of these two P10 ligands with five P20
(hydroxyethylamino)sulfonamide dipeptide isostere, several ligands that satisfy the substrate envelope constraints yielded
high-affinity moieties were identified for the P2 and P20 positions, ten inhibitors (Figure 1). The P20 moieties were selected based
including the bis-tetrahydrofurynyl (bis-THF) group in DRV on the previous structure-activity studies in the DRV series and
(Ghosh et al., 1998). However, most of these studies had the our computational library designs, and include 4-aminobenzene,
isobutyl group at the P10 position. We have identified other P10 4-methoxybenzene, 4-(hydroxymethyl)benzene, 1,3-benzodiox-
ligands capable of more extensive van der Waals contacts with olane, and benzothiazole (Altman et al., 2008; Ghosh et al., 2008;
Chemistry & Biology 20, 1116–1124, September 19, 2013 ª2013 Elsevier Ltd All rights reserved 1117

Chemistry & Biology
HIV-1 Protease Inhibitors to Avoid Drug Resistance
R2
O
a
b or c
R1 NH2
R2PhSO2Cl
S
BocHN
BocHN
N
H
R1
BocHN
N
O
80-90%
80-95%
O
OH
OH R1
5a-e
1
2a-b
R2 = CHO
3 R1 = (S)-2-methylbutyl
6a-e
7a-e
d
R2 = CH2OH
4 R1 = 2-ethyl-n-butyl
96%
NO2
R2
R2
O
O
O
H
H
O
O
O
g
S
S
O
O
O
N
H
N
RHN
N
O
O
O
O
75-98%
H
2 = NO2
H
OH R1
OH R1
R
h
R2 = NH2
9
R = Boc
R = H
R1 = (S)-2-methylbutyl
R1 = 2-ethyl-n-butyl
R
S
1 = ( )-2-methylbutyl
e
10a-e
111a-e
6a-e
93%
f
7a-e R1 = 2-ethyl-n-butyl
60%
6a,7a R2 = 4-NO2
10a, 11a
R2 = 4-NH2
OH
O
O
10b, 11b R2 = 4-OCH3
6b,7b R2 = 4-OCH3
6c,7c R2 = CH2OH
6d,7d R2 = 3,4-OCH2O-
H
H
10c, 11c R2 = CH2OH
10d, 11d
R
2 = 3,4-OCH2O-
8
Figure 2. Synthesis of Protease Inhibitors
Reagents and conditions: (a) R1NH2, EtOH, 80ꢀC, 3 hr; (b) aq. Na2CO3, CH2Cl2, 0ꢀC to room temperature (rt), overnight; (c) CH2Cl2, Et3N, 0ꢀC to rt, overnight; (d)
NaBH4, MeOH, 0ꢀC, 15 min; (e) TFA, CH2Cl2, 1 hr; (f) Py, p-NO2-PhOCOCl, 0ꢀC to rt, 24 hr; (g) DIEA, CH3CN, 0ꢀC to rt, 24 hr; (h) SnCl2.2H2O, EtOAc, 70ꢀC, 3 hr.
See Supplemental Experimental Procedures for more details.
Surleraux et al., 2005a, 2005b). The designed PIs were prepared With those caveats, some of these PIs have Ki values against
following the synthetic route illustrated in Figure 2. Briefly, ring WT protease an order of magnitude lower than the most potent
opening of chiral epoxide 1 with primary amines 2a–b provided FDA-approved drug DRV (Ki = 5 pM).
the amino alcohols 3 and 4. Reactions of sulfonyl chlorides
The PIs also retained potency against MDR variants of HIV-1
5a–e with 3 and 4 gave the intermediate (R)-(hydroxyethyla- protease. The Ki values for all PIs against the three MDR prote-
mino)sulfonamides 6–7. Boc deprotection followed by the reac- ases (M1, M2, and M3) were in the pM range. Most FDA-
tions of the free amines with bis-THF carbonate 9, prepared from approved PIs lose significant potency against M1 and M2,
chiral bis-THF alcohol 8, provided the target compound series except DRV and tipranavir (Figure 3; Table S2). All the PIs re-
10a–e and 11a–e.
tained potent activity against the M1 variant with Ki values com-
parable to or lower than DRV (M1 Ki = 25 pM), in particular three
PIs (10b, 11b, and 11c) had Ki values lower than DRV. Similarly,
most PIs inhibited the M2 variant with potency comparable to
The Designed Compounds Inhibit WT and
Drug-Resistant Protease Variants
The enzyme inhibition constants (Ki values) of the designed DRV. The M3 protease variant contains signature mutations of
inhibitors were determined against WT and three drug-resistant DRV resistance (I50V and A71V). Compared to WT protease,
HIV-1 protease variants, along with all nine FDA-approved PIs M3 susceptibility to DRV was 50-fold decreased, with a Ki of
(Figure 3; Table S2 available online). Two MDR protease variants 245 pM. In contrast, the designed inhibitors exhibited better po-
(M1: L10I/G48V/I54V/L63P/V82A; M2: L10I/L63P/A71V/G73S/ tency against M3, with Ki values consistently lower than that of
I84V/L90M) represent the pattern of resistance mutations that DRV, and as low as 6 pM for PIs 10a and 11d (Figure 3). Thus,
occur under the selective pressure of three or more currently this design strategy yielded highly potent PIs that retain enzyme
prescribed PIs in patients with HIV-1 infection (Wu et al., 2003). inhibition activity against MDR variants of HIV-1 protease.
The third variant (M3: I50V/A71V) is a signature-resistant variant
of APV/DRV, which shares the same scaffold used in the design larger, more flexible isopentyl and isohexyl groups resulted in
of the ten inhibitors. improved inhibition of drug-resistant protease variants. This is
Overall, replacement of the P10 isobutyl group of DRV with
All designed inhibitors were highly potent in inhibiting enzyme evident from better potency of PIs 10a and 11a, which incorpo-
activity, with Ki values in the 0.2–30 pM range against WT prote- rate the same 4-aminobenzene group at P20 as DRV. The modi-
ase. Five PIs had sub-pM Ki values (PIs 10c, 10d, 11b, 11c, and fications at the P10 position also proved to be advantageous
11d; Ki = 0.2–0.9 pM). Seven of the nine FDA-approved drugs are when combined with other groups at the P20 position. In partic-
two to three orders of magnitude weaker inhibitors of WT prote- ular, PIs 10c and 11c with the 4-(hydroxymethyl)benzene P20
ase (Ki = 46–284 pM), while two are low picomolar binders (LPV group were highly potent inhibitors of WT protease and retained
and DRV, Ki = 5 pM). The precise measurements of Ki values in low pM activity against resistant variants. Thus, incorporation of
the low pM range using standard biochemical assays are limited optimized ligands that can retain interactions with drug-resis-
ꢀ
(Gulnik and Eissenstat, 2008; Kuzmic, 1996; Miller et al., 2006). tance protease variants coupled with the substrate envelope
This makes direct comparison of these PIs with DRV, and each constraint proved to be a useful strategy for designing robust
other, challenging, as all have Ki values in the low pM range. inhibitors.
1118 Chemistry & Biology 20, 1116–1124, September 19, 2013 ª2013 Elsevier Ltd All rights reserved

Chemistry & Biology
HIV-1 Protease Inhibitors to Avoid Drug Resistance
Figure 3. Binding Affinities of APV, DRV,
and the Ten Designed Protease Inhibitors
to WT and Drug-Resistant Variants of
HIV-1 Protease
Inhibitory activity was determined by a FRET-
based enzymatic assay; Ki values are the average
of at least three independent measurements.
See also Table S2.
Improved Antiviral Activity against
WT and Drug-Resistant HIV Strains
The PIs were tested for antiviral activity
against WT HIV from clades A, B, and C,
as well as a diverse panel of 12 clinically
relevant patient-derived drug-resistant
viruses using PhenoSense HIV assays
The Inhibitors Fit within the Substrate Envelope with
Enhanced Protease Contacts
(Monogram Biosciences, San Francisco, CA) and compared
with DRV (Figure 5; Table S3; Figure S1). The average half-
The crystal structures of the ten potent inhibitors were deter- maximal effective concentration (EC50) values against all 12
mined in complex with WT HIV-1 protease. The crystals viral variants tested and fold changes in potency with respect
˚
diffracted to 1.45–1.95 A, yielding high-resolution complex to WT strain were used to evaluate PI resistance profiles
structures. The crystallographic and refinement statistics, as (Table S3).
well as Protein Data Bank (PDB) accession codes for all of the
structures are listed in Table S1.
All ten inhibitors exhibited subnanomolar potency against WT
virus from three different clades (EC50 = 0.04–0.4 nM). PIs 10b,
The co-crystal structures of all ten PIs in complex with the 10d, and 11b are more potent against WT clade B virus
protease superimpose with each other extremely well (Fig- compared to DRV (EC50 = 0.4 nM).
ure 1B). The protease backbone is very similar in all the com-
The designed PIs even retained antiviral potency against
plexes, with a root-mean-square deviation (rmsd) of 0.11– MDR strains, further validating the enzymatic assays. The
˚
0.16 A for Ca atoms. Even the protease side chains, including average EC50 values against 12 different drug-resistant strains
those in the active site, have similar conformations in all were lower than those for DRV for all ten PIs (Table S3; Fig-
the crystal structures, except minor local changes in side ure 5). Overall, the PIs were 1.5–4.5 times more potent than
chain conformations surrounding the different P10 and P20 DRV against drug-resistant HIV strains tested. There was no
groups. These tight-binding PIs ‘‘lock’’ into the active site and significant difference in the EC50 values of PIs with the isopentyl
induce a tightly shared rigid protease conformation with little (10a–d) versus the isohexyl (11a–d) group at the P10 position.
variability.
However, the P20 phenylsulfonamide groups appear to have
Except for the chemically diverse P10 and P20 groups, the more effect on the potency against drug-resistant strains.
inhibitors superimpose very well in all the crystal structures Thus, consistent with previous SAR results in the DRV series
and fit well within the substrate envelope (Figure 1C). Despite (Surleraux et al., 2005a), compounds with the 4-methoxyben-
inherent flexibility, the P10 isopentyl and isohexyl groups adopt zene (10b and 11b) and 1,3-benzodioxolane (10d and 11d)
the same conformation in all the complexes, similar to the isobu- groups at the P20 position exhibited better potency than the
tyl group in DRV bound to protease (Figure 4A). The key differ- corresponding 4-aminobenzene analogs 10a and 11a. How-
ence is the enhanced van der Waals contacts of the longer ever, in contrast to the corresponding DRV analog (Ghosh
isopentyl (displayed for PI 10a in Figure 4B) and isohexyl (dis- et al., 2008), PIs with the 4-(hydroxymethyl)benzene P20 group
played for PI 11a in Figure 4C) groups within the S10 pocket of (10c and 11c) also exhibited highly potent activity, likely due
the protease active site, compared to those of the isobutyl group to the increased lipophilicity. In a previous study, HIV PIs
of DRV. The P10 group makes vdW interactions with the hydro- with cLogP values in the 3.7–5.0 range exhibited antiviral
phobic residues I50, P81, V82, and I84 in the active site of the potencies against drug-resistant HIV variants better than DRV
protease. With the exception of P81, these residues are key sites (cLogP = 2.88; Miller et al., 2006). Because the cLogP values
of multidrug resistance, and mutate to those with smaller side of the PIs are all in the desired range (Table S4), the
chains (Ile to Val, Val to Ala) to confer resistance, such as in increased lipophilicities, in addition to other factors, may be
the case of I50V mutation in the DRV signature-resistance variant contributing to the improved antiviral potencies compared to
M3. The isopentyl and isohexyl groups in the PIs fill the S10 those of DRV.
pocket better compared to the isobutyl group of DRV, and their
DRV retains low nanomolar activity against most MDR strains,
flexibility makes them capable of adapting and forming van der but loses significant potency against two MDR strains with the
Waals contacts even with the mutated shorter side chain resi- I50V protease mutation. However, a number of the PIs retain bet-
dues. These enhanced contacts are consistent with retained ter potency against these MDR strains, particularly compounds
affinities to drug-resistant HIV protease variants with mutations 10b and 10d with the 4-methoxybenzene and 1,3-benzodioxo-
around the S10 pocket (Figure 3; Table S2).
lane groups at the P20 position. In fact, compound 10b retains
Chemistry & Biology 20, 1116–1124, September 19, 2013 ª2013 Elsevier Ltd All rights reserved 1119

Chemistry & Biology
HIV-1 Protease Inhibitors to Avoid Drug Resistance
Figure 4. Comparison of van der Waals
Contacts with the Protease for P10 Groups
in PIs
(A) Superposition of DRV, PI 10a, and PI 11a
structures bound to WT HIV-1 protease.
(B) Isobutyl P10 group in DRV.
(C and D) Isopentyl P10 group in PI 10a (C), iso-
hexyl P10 group in PI 11a (D). The longer P10
groups make enhanced van der Waals contacts
with the protease.
in vitro pharmacokinetic assays. The
aqueous solubility, plasma protein bind-
ing, plasma stability, and microsomal
stability in rat and human liver micro-
somes of all PIs were determined by
Apredica (Watertown, MA) using DRV as
a control (Tables 1 and S4).
The oral bioavailability of
a drug
depends in part on water solubility
and lipophilicity. The latter is estimated
by the calculated partition coefficient
cLogP, which is optimized to balance
between aqueous solubility and transport
through the hydrophobic cell membrane
(Miller et al., 2006). Compounds with
cLogP values in the 3.5–5.0 range have
been reported to have the best antiviral
activity as HIV PIs (Miller et al., 2006).
The cLogP values of the PIs are all in the
desired range (Table S4). In fact, they
have higher lipophilicities compared to
DRV, which may enhance cell penetra-
tion. Despite the enhanced lipophilicity,
the PIs also have good aqueous solubi-
lity except the PIs 10b and 11b with the
methoxybenzene group at the P20 posi-
tion. Overall, the increased lipophilicity
and reasonable aqueous solubility likely
contribute to the improved antiviral
activity profiles (EC50 values), which
depend on both the intrinsic binding
affinity to the target (Ki), and the transport
efficiency into the cell.
subnanomolar antiviral potency against 10 of the 12 MDR strains
The metabolic stability of PIs was evaluated by measuring the
tested and low nanomolar potency against the strains with the microsomal stability, plasma stability, and plasma protein bind-
I50V protease mutation, exhibiting the lowest average EC50 of ing (Tables 1 and S4). Assays were performed in rat and human
1.33 nM compared to 5.98 nM for DRV. Thus, the combination microsomes to determine the stability of PIs at 15-, 30-, and
of isopentyl group at P10 and 4-methoxybenzene group at P20 60-min time points and in the absence and presence of the
provided the compound (10b) with the best resistance profile.
CYP-450 inhibitor ritonavir (RTV; Table 1). Similar to DRV, RTV
The excellent resistance profiles of the PIs indicate that the boosting considerably enhanced the microsomal stability for
design strategy of filling the subsites without violating the sub- the majority of the PIs, which were otherwise not stable in the
strate envelope constraints was successful in optimizing the presence of liver microsomes at 60 min. Interestingly, two inhib-
inhibitor for robustness against resistance mutations.
itors with the benzothiazole P20 moiety (10e and 11e) showed
excellent stability in both rat and human microsomal tests
even without RTV boosting. Overall, the PIs have good pharma-
In Vitro Pharmacokinetic Evaluation
The exciting PhenoSense resistance profile data prompted cokinetic properties, encouraging their further development as
further evaluation of the PIs as potential drug candidates using drug candidates.
1120 Chemistry & Biology 20, 1116–1124, September 19, 2013 ª2013 Elsevier Ltd All rights reserved

Chemistry & Biology
HIV-1 Protease Inhibitors to Avoid Drug Resistance
Figure 5. Resistance Profiles of DRV and
the Ten PIs that Were Designed
Antiviral potencies (EC50 values) were obtained for
WT HIV from clades A, B, and C, and 12 different
patient-derived drug-resistant variants of the
virus. The PhenoSense HIV assays were per-
formed by Monogram Biosciences.
See also Table S3 and Figure S1.
The improved binding affinities of PIs
against drug-resistant protease variants
correlate with their exceptional antiviral
potencies against a panel of 12 patient-
derived drug resistant HIV strains in
PhenoSense HIV assays. The PIs retain
better antiviral potency against MDR
HIV variants less susceptible to DRV,
with EC50 values in the low nanomolar
range. However, MDR viruses insensitive
to DRV are also relatively less sensitive
to designed PIs, implying the same fac-
tors may render the inhibitors less effective in the antiviral
assays. The MDR viruses insensitive to DRV and the ten PIs
DISCUSSION
With the emergence of resistance, many drugs become obsolete contain a combination of protease secondary mutations outside
in the clinic against rapidly evolving disease targets, such as in the active site, which may interdependently confer resistance to
the case of HIV/AIDS. In HIV infection, error-prone reverse tran- all these inhibitors. Such secondary mutations may alter the
scriptase and the high replication rates of the virus result in a processivity of the protease to favor substrate processing over
diverse viral population, among which drug-resistant viral strains inhibitor binding, thereby decreasing susceptibility to inhibitors
are selected under the pressure of drug therapy. To avoid (Chang and Torbett, 2011; Muzammil et al., 2003; Xie et al.,
spending time and resources in developing drugs that rapidly 1999). Thus, the compensatory mutations outside the protease
become obsolete, strategies at the initial design stage need to active site may also be contributing toward the observed resis-
be implemented to minimize the chances of subsequent drug tance profiles.
resistance.
All ten inhibitor-protease complex structures superimpose
The substrate envelope model provides a framework for im- extremely well with each other, despite chemical diversity at
plementing drug resistance considerations into structure-based the P10 and P20 sites of the inhibitors. The exceptionally high
drug design. For drug resistance to occur, the target needs to be affinity of the designed sub- and low-picomolar inhibitors may
able to carry out its biologic function while at the same time be locking the protease structure and not allowing much confor-
avoiding drug binding. The chances of this happening are dimin- mational variability. Future studies of PI complex dynamics will
ished if the designed drug binds the target similarly to the natural provide insights into the effect of sub-picomolar binders on the
substrates. Protrusions beyond the substrate envelope cause protease conformational and dynamic behavior.
vulnerable sites for drug-resistance mutations (King et al.,
In conclusion, in quickly evolving therapeutic targets, restrict-
2004a) because mutations into bulkier side chains impact drug ing inhibitors to fit within the substrate envelope is a key factor in
binding without affecting the substrates.
minimizing the susceptibility to drug resistance.
The results presented here demonstrate that optimally filling
the substrate envelope is essential to developing potent inhibi- SIGNIFICANCE
tors. In cases where the designed drug occupies a smaller vol-
ume than the substrates, mutations into shorter side chains Overcoming resistance is crucial in designing robust drugs
may diminish van der Waals interactions with the inhibitor, while against HIV/AIDS. Resistance to direct-acting antiviral
still maintaining sufficient interactions with the larger natural sub- drugs, including HIV-1 protease inhibitors, occurs when
strates. In the DRV signature-resistant HIV-1 protease variant, the target mutates to avoid drug binding, while maintaining
the I50V mutation causes loss of contacts with the P10 isobutyl activity on substrates. Our design strategy of optimizing in-
group. The longer isopentyl and isohexyl moieties introduced hibitor fit within the substrate envelope yielded high-affinity
here to the P10 site optimally fill the substrate envelope and inhibitors of multidrug resistant HIV protease variants,
have flexibility, which enables sustaining van der Waals interac- which exhibited improved antiviral potencies as compared
tions even when resistant mutations occur. Thus, the inhibitors to DRV against a large panel of drug-resistant HIV strains.
have improved van der Waals contacts with the protease while Some of these potent inhibitors also have promising phar-
still staying within the substrate envelope and retain potency macokinetic profiles, even in the absence of a boosting
against a wider variety of HIV-1 protease mutants.
agent, encouraging their potential development as drugs.
Chemistry & Biology 20, 1116–1124, September 19, 2013 ª2013 Elsevier Ltd All rights reserved 1121

Chemistry & Biology
HIV-1 Protease Inhibitors to Avoid Drug Resistance
Table 1. Microsomal Stability Profile of Protease Inhibitors
Microsomal Stability (avg. % fraction remaining)
Microsomal CLint
(ml/min/mg)
Microsomal T1/2 (min)
at 30 min
at 60 min
Compound
10a
RTV
ꢁ
+
Human
42
81
0
Rat
24
51
0
Human
19
66
0
Rat
9
Human
81
Rat
129
67
Human
26
Rat
16
35
0
21
99
31
2
10b
10c
10d
10e
11a
11b
11c
11d
11e
DRV
ꢁ
+
776
19
885
518
664
79
3
79
0
0
68
0
0
110
4
4
ꢁ
+
0
0
590
8
3
93
9
45
0
89
4
22
0
>180
9
26
2
ꢁ
+
232
24
987
771
53
80
32
86
50
87
0
0
60
33
62
26
71
0
0
85
3
ꢁ
+
64
71
13
30
18
50
0
31
74
2
82
25
40
106
10
18
14
28
2
21
20
99
ꢁ
+
69
204
117
150
74
30
11
19
20
0
16
126
2
ꢁ
+
844
23
75
0
66
0
90
ꢁ
+
829
26
907
27
3
64
0
79
0
69
0
57
0
81
78
3
ꢁ
+
617
46
712
156
9
3
55
100
96
37
100
21
84
62
40
84
46
83
95
13
100
4
45
13
>180
48
25
103
ꢁ
+
87
44
23
65
7
>180
>180
21
3
44
ꢁ
+
101
0
83
20
>180
RTV, ritonavir; CLint, microsomal intrinsic clearance; T1/2, half-life.
See also Table S4.
hundred eighty frames were collected per crystal with an angular separation
of 1ꢀ and no overlap between frames. The data processing of the frames
was carried out using the programs DENZO and ScalePack, respectively
(Minor, 1993; Otwinowski et al., 1997). Intensity data for 10c, 22a, and 11e
were collected under cryogenic conditions at the BioCARS 14-BMC beamline
at Argonne National Laboratory (Advanced Photon Source, Argonne, IL).
Diffraction images were indexed and scaled using the program HKL2000
(Otwinowski et al., 1997).
This study provides strong rationale for incorporating sub-
strate envelope constraints in the initial stages of drug
design as a powerful strategy to design potent inhibitors
that evade resistance. Drug resistance extends well beyond
HIV, thwarting the success of many drugs to quickly evolving
targets in pathogens and cancer. This strategy is broadly
applicable to the design of inhibitors against such quickly
evolving disease targets, facilitating a more rational and
rapid selection of drug candidates that are robust and thus
less susceptible to resistance.
The crystal structures were solved and refined with the programs within the
CCP4 interface (CCP4, 1994). Structure solutions for all the WT PI complexes
were obtained with the molecular replacement package AMoRe (Navaza,
1994) using 1F7A (Prabu-Jeyabalan et al., 2000) as the starting model. Upon
obtaining solution, the molecular replacement phases were further improved
using ARP/wARP (Morris et al., 2002) to build solvent molecules into the unac-
counted regions of electron density. Model building was performed using the
interactive graphics program Coot (Emsley and Cowtan, 2004). Conjugate
gradient refinement using Refmac5 (Murshudov et al., 1997) was performed
by incorporating Schomaker and Trueblood tensor formulation of TLS (trans-
lation, libration, screw-rotation) parameters (Kuriyan and Weis, 1991; Scho-
EXPERIMENTAL PROCEDURES
Protein Crystallography
The expression, isolation, and purification of wild-type and mutant HIV-1 pro-
teases used for crystallization and binding experiments were carried out as
previously described (King et al., 2002). Cocrystals of the inhibitors with the
WT protease were grown at room temperature by hanging drop vapor diffusion
method. Protease concentration of 1.6 mg/ml with 3-fold molar excess of in-
hibitors was used to set the crystallization drops with the reservoir solution
consisting of 126 mM phosphate buffer at pH 6.2, 63 mM sodium citrate,
and 24%–29% ammonium sulfate.
maker and Trueblood, 1968; Tickle and Moss, 1999). The working R (Rfactor
)
and its cross-validation (Rfree) were monitored throughout the refinement.
The data collection and refinement statistics are shown in Table S5.
ACCESSION NUMBERS
The crystals used for data collection were mounted in Mitegen Micromounts
and flash-frozen over a nitrogen stream. Intensity data for complex structures
of 10a, 10b, 10d, 10e, 11b, 11c, and 11d were collected at ꢁ80ꢀC on an in-
house Rigaku X-ray generator equipped with an R-axis IV image plate. One
The PDB accession numbers for the HIV-1 protease-inhibitor complexes
reported in this paper are 3O99, 3O9A, 3O9B, 3O9C, 3O9D, 3O9E, 3O9F,
3O9G, 3O9H, and 3O9I.
1122 Chemistry & Biology 20, 1116–1124, September 19, 2013 ª2013 Elsevier Ltd All rights reserved

Chemistry & Biology
HIV-1 Protease Inhibitors to Avoid Drug Resistance
SUPPLEMENTAL INFORMATION
Ghosh, A.K., Leshchenko-Yashchuk, S., Anderson, D.D., Baldridge, A.,
Noetzel, M., Miller, H.B., Tie, Y., Wang, Y.F., Koh, Y., Weber, I.T., and
Mitsuya, H. (2009). Design of HIV-1 protease inhibitors with pyrrolidinones
and oxazolidinones as novel P10-ligands to enhance backbone-binding
interactions with protease: synthesis, biological evaluation, and protein-ligand
X-ray studies. J. Med. Chem. 52, 3902–3914.
Supplemental Information includes Supplemental Experimental Procedures,
one figure, and four tables and can be found with this article online at http://
dx.doi.org/10.1016/j.chembiol.2013.07.014.
ACKNOWLEDGMENTS
Ghosh, A.K., Chapsal, B.D., Baldridge, A., Steffey, M.P., Walters, D.E., Koh, Y.,
Amano, M., and Mitsuya, H. (2011). Design and synthesis of potent HIV-1 pro-
tease inhibitors incorporating hexahydrofuropyranol-derived high affinity P(2)
ligands: structure-activity studies and biological evaluation. J. Med. Chem.
54, 622–634.
This work was made possible by grants from the National Institute of General
Medical Sciences, National Institutes of Health (P01-GM66524, AI41404,
AI43198, GM065418, and GM082209) and ARRA Supplement
(P01GM066524-08S1). The authors would like to thank Dr. Norton Peet and
Prof. William Royer for helpful discussions and Rajintha Bandaranayake,
Seema Mittal, and Keith Romano for collecting some of the data at BioCARS
beamline of the Advanced Photon Source, Argonne National Laboratory. Use
of the Advanced Photon Source for X-ray data collection was supported by the
US Department of Energy, Basic Energy Sciences, Office of Science under
contract no. DE-AC02-06CH11357. Use of the BioCARS Sector 14 was sup-
ported by the National Institutes of Health, National Center for Research
Resources under grant number RR007707. We also thank Kaneka USA for
generous gifts of chiral epoxides.
Gulnik, S.V., and Eissenstat, M. (2008). Approaches to the design of HIV
protease inhibitors with improved resistance profiles. Curr Opin HIV AIDS 3,
633–641.
Hogg, R.S., Heath, K.V., Yip, B., Craib, K.J.P., O’Shaughnessy, M.V.,
Schechter, M.T., and Montaner, J.S.G. (1998). Improved survival among
HIV-infected individuals following initiation of antiretroviral therapy. JAMA
279, 450–454.
Kempf, D.J., Marsh, K.C., Kumar, G., Rodrigues, A.D., Denissen, J.F.,
McDonald, E., Kukulka, M.J., Hsu, A., Granneman, G.R., Baroldi, P.A., et al.
(1997). Pharmacokinetic enhancement of inhibitors of the human immunode-
ficiency virus protease by coadministration with ritonavir. Antimicrob. Agents
Chemother. 41, 654–660.
Received: March 27, 2013
Revised: July 12, 2013
Accepted: July 23, 2013
Published: September 5, 2013
King, N.M., Melnick, L., Prabu-Jeyabalan, M., Nalivaika, E.A., Yang, S.S., Gao,
Y., Nie, X., Zepp, C., Heefner, D.L., and Schiffer, C.A. (2002). Lack of synergy
for inhibitors targeting a multi-drug-resistant HIV-1 protease. Protein Sci. 11,
418–429.
REFERENCES
King, N.M., Prabu-Jeyabalan, M., Nalivaika, E.A., and Schiffer, C.A. (2004a).
Combating susceptibility to drug resistance: lessons from HIV-1 protease.
Chem. Biol. 11, 1333–1338.
Ali, A., Reddy, G.S.K.K., Cao, H., Anjum, S.G., Nalam, M.N.L., Schiffer, C.A.,
and Rana, T.M. (2006). Discovery of HIV-1 protease inhibitors with picomolar
affinities incorporating N-aryl-oxazolidinone-5-carboxamides as novel P2
ligands. J. Med. Chem. 49, 7342–7356.
King, N.M., Prabu-Jeyabalan, M., Nalivaika, E.A., Wigerinck, P., de Be´ thune,
M.P., and Schiffer, C.A. (2004b). Structural and thermodynamic basis for the
binding of TMC114, a next-generation human immunodeficiency virus type 1
protease inhibitor. J. Virol. 78, 12012–12021.
Ali, A., Reddy, G.S.K.K., Nalam, M.N.L., Anjum, S.G., Cao, H., Schiffer, C.A.,
and Rana, T.M. (2010). Structure-based design, synthesis, and structure-
activity relationship studies of HIV-1 protease inhibitors incorporating phenyl-
oxazolidinones. J. Med. Chem. 53, 7699–7708.
Kuriyan, J., and Weis, W.I. (1991). Rigid protein motion as a model for crystal-
lographic temperature factors. Proc. Natl. Acad. Sci. USA 88, 2773–2777.
Altman, M.D., Ali, A., Reddy, G.S.K.K., Nalam, M.N.L., Anjum, S.G., Cao, H.,
Chellappan, S., Kairys, V., Fernandes, M.X., Gilson, M.K., et al. (2008). HIV-1
protease inhibitors from inverse design in the substrate envelope exhibit subna-
nomolar binding to drug-resistant variants. J. Am. Chem. Soc. 130, 6099–6113.
ꢀ
Kuzmic, P. (1996). Program DYNAFIT for the analysis of enzyme kinetic data:
application to HIV proteinase. Anal. Biochem. 237, 260–273.
Miller, J.F., Andrews, C.W., Brieger, M., Furfine, E.S., Hale, M.R., Hanlon,
M.H., Hazen, R.J., Kaldor, I., McLean, E.W., Reynolds, D., et al. (2006).
Ultra-potent P1 modified arylsulfonamide HIV protease inhibitors: the discov-
ery of GW0385. Bioorg. Med. Chem. Lett. 16, 1788–1794.
Cai, Y., and Schiffer, C.A. (2010). Decomposing the energetic impact of drug
resistant mutations in HIV-1 protease on binding DRV. J. Chem. Theory
Comput. 6, 1358–1368.
CCP4 (Collaborative Computational Project, Number 4). (1994). The CCP4
Minor, W. (1993). XdisplayF (West Lafayette, Indiana: Purdue University).
suite: programs for protein crystallography. Acta Crystallogr.
D Biol.
Morris, R.J., Perrakis, A., and Lamzin, V.S. (2002). ARP/wARP’s model-build-
ing algorithms. I. The main chain. Acta Crystallogr. D Biol. Crystallogr. 58,
968–975.
Crystallogr. 50, 760–763.
Chang, M.W., and Torbett, B.E. (2011). Accessory mutations maintain stability
in drug-resistant HIV-1 protease. J. Mol. Biol. 410, 756–760.
Murshudov, G.N., Vagin, A.A., and Dodson, E.J. (1997). Refinement of macro-
molecular structures by the maximum-likelihood method. Acta Crystallogr. D
Biol. Crystallogr. 53, 240–255.
Chellappan, S., Kiran Kumar Reddy, G.S., Ali, A., Nalam, M.N.L., Anjum, S.G.,
Cao, H., Kairys, V., Fernandes, M.X., Altman, M.D., Tidor, B., et al. (2007).
Design of mutation-resistant HIV protease inhibitors with the substrate
envelope hypothesis. Chem. Biol. Drug Des. 69, 298–313.
Muzammil, S., Ross, P., and Freire, E. (2003). A major role for a set of non-
active site mutations in the development of HIV-1 protease drug resistance.
Biochemistry 42, 631–638.
Cihlar, T., He, G.X., Liu, X., Chen, J.M., Hatada, M., Swaminathan, S.,
McDermott, M.J., Yang, Z.Y., Mulato, A.S., Chen, X., et al. (2006).
Suppression of HIV-1 protease inhibitor resistance by phosphonate-mediated
solvent anchoring. J. Mol. Biol. 363, 635–647.
Nalam, M.N.L., and Schiffer, C.A. (2008). New approaches to HIV protease
inhibitor drug design II: testing the substrate envelope hypothesis to avoid
drug resistance and discover robust inhibitors. Curr Opin HIV AIDS 3, 642–646.
Emsley, P., and Cowtan, K. (2004). Coot: model-building tools for molecular
Nalam, M.N.L., Ali, A., Altman, M.D., Reddy, G.S.K.K., Chellappan, S., Kairys,
V., Ozen, A., Cao, H., Gilson, M.K., Tidor, B., et al. (2010). Evaluating the sub-
strate-envelope hypothesis: structural analysis of novel HIV-1 protease inhib-
itors designed to be robust against drug resistance. J. Virol. 84, 5368–5378.
graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132.
Ghosh, A.K., Krishnan, K., Walters, D.E., Cho, W., Cho, H., Koo, Y., Trevino, J.,
Holland, L., and Buthod, J. (1998). Structure based design: novel spirocyclic
ethers as nonpeptidal P2-ligands for HIV protease inhibitors. Bioorg. Med.
Chem. Lett. 8, 979–982.
Navaza, J. (1994). AMoRe: An automated package for molecular replacement.
Acta Crystallogr. A 50, 157–163.
Ghosh, A.K., Chapsal, B.D., Weber, I.T., and Mitsuya, H. (2008). Design of HIV
protease inhibitors targeting protein backbone: an effective strategy for
combating drug resistance. Acc. Chem. Res. 41, 78–86.
Otwinowski, Z., Minor, W., and Carter, C.W., Jr. (1997). Processing of X-ray
diffraction data collected in oscillation mode. In Methods in Enzymology,
Chemistry & Biology 20, 1116–1124, September 19, 2013 ª2013 Elsevier Ltd All rights reserved 1123

Chemistry & Biology
HIV-1 Protease Inhibitors to Avoid Drug Resistance
C.W.J. Carter and R.M. Sweet, eds. (New York: Academic Press),
pp. 307–326.
Surleraux, D.L., de Kock, H.A., Verschueren, W.G., Pille, G.M., Maes, L.J.,
Peeters, A., Vendeville, S., De Meyer, S., Azijn, H., Pauwels, R., et al.
(2005a). Design of HIV-1 protease inhibitors active on multidrug-resistant
virus. J. Med. Chem. 48, 1965–1973.
Palella, F.J., Jr., Delaney, K.M., Moorman, A.C., Loveless, M.O., Fuhrer, J.,
Satten, G.A., Aschman, D.J., Holmberg, S.D., and The, H.I.V.O.S.I.; HIV
Outpatient Study Investigators. (1998). Declining morbidity and mortality
among patients with advanced human immunodeficiency virus infection.
N. Engl. J. Med. 338, 853–860.
Surleraux, D.L., Tahri, A., Verschueren, W.G., Pille, G.M., de Kock, H.A.,
Jonckers, T.H., Peeters, A., De Meyer, S., Azijn, H., Pauwels, R., et al.
(2005b). Discovery and selection of TMC114, a next generation HIV-1 protease
inhibitor. J. Med. Chem. 48, 1813–1822.
Prabu-Jeyabalan, M., Nalivaika, E., and Schiffer, C.A. (2000). How does a
symmetric dimer recognize an asymmetric substrate? A substrate complex
of HIV-1 protease. J. Mol. Biol. 301, 1207–1220.
Tickle, I.J., and Moss, D.S. (1999). Modeling rigid-body thermal motion in
macromolecular crystal structure refinement. In IUCr99 Computing
SchoolIUCr99 Computing School (London: IUCr).
Prabu-Jeyabalan, M., Nalivaika, E., and Schiffer, C.A. (2002). Substrate shape
determines specificity of recognition for HIV-1 protease: analysis of crystal
structures of six substrate complexes. Structure 10, 369–381.
Wensing, A.M., van Maarseveen, N.M., and Nijhuis, M. (2010). Fifteen years of
HIV Protease Inhibitors: raising the barrier to resistance. Antiviral Res. 85,
59–74.
Romano, K.P., Ali, A., Royer, W.E., and Schiffer, C.A. (2010). Drug resistance
against HCV NS3/4A inhibitors is defined by the balance of substrate recogni-
tion versus inhibitor binding. Proc. Natl. Acad. Sci. USA 107, 20986–20991.
Wu, T.D., Schiffer, C.A., Gonzales, M.J., Taylor, J., Kantor, R., Chou, S.,
Israelski, D., Zolopa, A.R., Fessel, W.J., and Shafer, R.W. (2003). Mutation pat-
terns and structural correlates in human immunodeficiency virus type 1 prote-
ase following different protease inhibitor treatments. J. Virol. 77, 4836–4847.
¨
Romano, K.P., Ali, A., Aydin, C., Soumana, D., Ozen, A., Deveau, L.M., Silver,
C., Cao, H., Newton, A., Petropoulos, C.J., et al. (2012). The molecular basis of
drug resistance against hepatitis C virus NS3/4A protease inhibitors. PLoS
Pathog. 8, e1002832.
Xie, D., Gulnik, S., Gustchina, E., Yu, B., Shao, W., Qoronfleh, W., Nathan, A.,
and Erickson, J.W. (1999). Drug resistance mutations can effect dimer stability
of HIV-1 protease at neutral pH. Protein Sci. 8, 1702–1707.
Schomaker, V., and Trueblood, K.N. (1968). On the rigid-body motion of mol-
Youle, M. (2007). Overview of boosted protease inhibitors in treatment-expe-
ecules in crystals. Acta Crystallogr. B 24, 63–76.
rienced HIV-infected patients. J. Antimicrob. Chemother. 60, 1195–1205.
´
´
´
Stranix, B.R., Sauve, G., Bouzide, A., Cote, A., Sevigny, G., and Yelle, J. (2003).
Lysine sulfonamides as novel HIV-protease inhibitors: optimization of the
Nepsilon-acyl-phenyl spacer. Bioorg. Med. Chem. Lett. 13, 4289–4292.
Zeldin, R.K., and Petruschke, R.A. (2004). Pharmacological and therapeutic
properties of ritonavir-boosted protease inhibitor therapy in HIV-infected
patients. J. Antimicrob. Chemother. 53, 4–9.
1124 Chemistry & Biology 20, 1116–1124, September 19, 2013 ª2013 Elsevier Ltd All rights reserved